Pilot Baseline
~$10k one time
Two weeks of sentinel tests on 5–10 hospital feeds.
- Time-of-day latency heat map
- Silent-feed and MPI findings
- Top-10 fix list
- One readout session
Regier HealthIT · Assessment demo
Did the right alert reach the right care coordinator, on time? Most HIEs can’t say for sure.
An encounter notification can fail quietly at five places between the hospital’s ADT feed and the care coordinator’s worklist. A feed goes silent over a weekend. A discharge sits in an MPI review queue. A Direct address bounces. Nobody gets paged, and nobody knows until a patient is readmitted without a follow-up call.
You promised your hospitals compliance. Can you prove it? Many HIEs send the encounter notifications their member hospitals rely on for the CMS e-notification Condition of Participation, and HIEs are expected to meet 21st Century Cures Act information-blocking expectations. Few can show the evidence. The Health Check produces dated timing and delivery evidence behind that promise.
● Demo · synthetic data for a fictional HIE
The blind spot
Interface monitoring tells you a channel is up. It doesn’t tell you whether a discharge from a rural hospital at 11:40 PM on a Saturday reached the right primary care practice before Monday.
Silent feeds
A hospital interface stops sending A03 discharges after a weekend change. Volume drops to zero, but nothing errors, so nothing alarms.
Stuck matches
Missing MRNs or demographic drift push messages into MPI manual review. The alert waits in a queue, or never fires.
Unproven delivery
Portal, EHR and Direct deliveries go out without confirmation tracking. Stale Direct addresses bounce and nobody is told.
Why now
In “Health Data Quality Across Three HIEs Nationwide” (Civitas Networks for Health case study, published as “Measuring What Matters,” November 16, 2026), three member HIEs describe the same shift:
Site-level scorecards, dashboards, data traffic monitoring, and automated validation against HL7 and C-CDA specifications, with feedback sent straight to participants.
Measures aligned to USCDI, PIQI, HEDIS, Medicaid quality programs and NCQA’s Data Aggregator Validation (DAV) program.
Quality councils and committees that set onboarding thresholds by care setting, so data quality becomes an operational discipline, not a side project.
Reducing EMPI duplicates and strengthening matching, because duplicate records point to wider downstream problems.
Much of that work is being designed and built in-house, over years. The ENS Alert Health Check is a turnkey, ENS-focused way to get there: continuous monitoring, scorecards and evidence for the alert pipeline, starting with a two-week baseline instead of a multi-year build.
Summary of a public case study. Civitas Networks for Health has not reviewed or endorsed this service.
Still open in the center
How it works
Each test follows one alert through the whole pipeline and records when it passed each checkpoint, and where it stopped.
Admit, discharge and update events leave the hospital EHR.
A01A03A08Checks: arrival time, A03 coverage, PV1-36
Messages are parsed, routed and filtered.
MirthOpen Integration EngineChecks: errors, filters, queue time
The patient is linked to the right person in the master patient index.
Auto-matchReview queueChecks: match rate, review delay
The matched patient is checked against each subscriber’s attributed roster.
Care teamPayer panelChecks: roster hit, suppression
The alert goes out to the subscriber’s channel.
EHR inboxDirect / DirectTrustPortalChecks: sent, bounced, delivery time
The receiving system confirms receipt, and a coordinator opens it.
ACK / MDNOpenedChecks: confirmation, time to open
Clearly tagged test patients are enrolled on a test panel. On a schedule, the Health Check injects A01, A03 and A08 messages for them through each hospital’s feed and times every checkpoint.
Each filled hour is a test run for one hospital feed. Denser coverage during the discharge rush and the overnight batch window.
Interactive sample dashboard
Every number here is generated from a seeded random model of a fictional HIE with 12 fictional hospitals. Change the filters and the charts recompute.
Median and 95th-percentile minutes at each checkpoint, sentinel runs.
Median minutes from hospital event to confirmed delivery. Hover or tap a cell.
Sort by any column. Select a hospital to see its field completeness and top issues.
Hourly inbound ADT volume against the feed’s normal rhythm.
Prairie State HIE against anonymized peer HIEs A–F (sample).
Alerts delivered by Direct Secure Messaging over the DirectTrust network. Addresses are checked against the DirectTrust directory, and delivery notifications (MDN) are tracked for each message.
Full weeks ending Friday: confirmed, delayed (confirmation after 1 hour), bounced, and no confirmation.
Subscriber Direct addresses by directory status.
CoP evidence when a hospital’s alerts are delivered by Direct.
MPI matching rules are often tuned on the most common naming patterns. Synthetic sentinel cohorts built around local naming conventions show where matching underperforms for specific communities, so the alerts reach those patients’ care teams too. Cohorts reflect naming and record conventions only.
Share of sentinel messages linked automatically to the right person. Dashed line: baseline cohort.
Share of sentinel messages that created a new record or went unmatched although the person already existed.
Synthetic test sets. Share these with your MPI vendor’s team; the tuning itself is done in the MPI.
Share of A01 and A03 sentinel alerts delivered within each band.
Each hospital’s evidence packet would include the run log, timestamps at each checkpoint, and the delivery confirmation record for every sentinel alert in the period.
Synthetic sentinel patients carrying opt-out, restricted, or 42 CFR Part 2 flags are pushed through the pipeline. Each scenario checks that data is blocked or sequestered from queries, that alerts are suppressed or allowed according to the HIE’s policy, and that an audit trail is written.
Pass means the system behaved as the HIE’s policy expects for that scenario. Select a cell for details.
Select a cell to see the test detail.
Median and 95th-percentile hours from opt-out recorded to enforced, by system. Target line: 24 hours (illustrative; set by HIE policy).
For HIEs connected to Carequality, eHealth Exchange, or TEFCA through a QHIN: synthetic test queries and subscriptions, timed and scored.
Median seconds for patient discovery plus document query, by network.
Sample metrics for the selected period.
Public-health reporting from the same hospitals: syndromic surveillance (NSSP-style timeliness and completeness), electronic case reporting (eICR and Reportability Response), and electronic lab reporting (ELR) to the state. Sample data, not an official CDC or state report.
Share of visits whose first message arrived within 24 and 48 hours.
Share of visits with each data element populated.
eICRs sent, received by the public health routing service, and Reportability Responses (RR) returned, by week.
Per-hospital readiness for public-health measures (sample).
Sample data only. Hospital names, peer HIEs and every value are fictional and generated in your browser from a fixed seed.
Sample findings
A Full Assessment ends with findings your board can read and a fix list your interface team can work.
Packages and pricing
Starting-at prices and typical ranges. Final scope depends on the number of hospital feeds, delivery channels and networks.
~$10k one time
Two weeks of sentinel tests on 5–10 hospital feeds.
$15k–$25k one time
Every hospital feed, end to end.
$2k–$4k per month
Continuous monitoring after the baseline.
Prices are starting points and typical ranges, not quotes. Travel, if any, is billed separately.
Next step
Tell me how many hospital feeds you run, which engine and MPI you use, and how alerts reach subscribers. I’ll come back with a scoped Pilot Baseline.